Why the demo failed the test:
- Appears to be stuck at console, does not render graphics

Why the demo failed the test:
- Bochs doesn't like the "PC speaker speech" of this demo, and will hang/slow down

Why the demo failed the test:
- QEMU does not render PC speaker output

Why the demo failed the test:
- Have not yet figured out if Bochs can emulate the PC speaker playback in this demo

Why the demo failed the test:
- QEMU does not render PC speaker output

Why the demo failed the test:
- Doesn't like Bochs SB16 emulation

Why the demo failed the test:
- Hangs on startup (SB16 emulation given DSP command it doesn't understand)

Why the demo failed the test:
- hpel panning is jumpy

Why the demo failed the test:
- Can't tell if hpel panning correct, runs way too fast

Why the demo failed the test:
- Disney Sound Source / LPT DAC output does not work

Why the demo failed the test:
- Disney Sound Source / LPT DAC output does not work

Why the demo failed the test:
- Disney Sound Source / LPT DAC output does not work

Why the demo failed the test:
- Refresh rate issues
- Some parts do not render
- Some mode changes cause Bochs to hang for a significant amount of time
The following are not expected of Bochs:
- No LPT DAC support

Why the demo failed the test:
- Runs way too fast
- Some parts do not render at all
- Per-scanline effects not supported
Not QEMU's fault:
- Lack of LPT DAC emulation

Why the demo failed the test:
- Scrollers run way too fast
- End PC speaker music inaudible
- End screen scroll is too slow

Why the demo failed the test:
- Scroller runs way too fast
- PC speaker music inaudible

Why the demo failed the test:
- Does not like Bochs SB16 emulation

Why the demo failed the test:
- Does not like SB16 emulation

Why the demo failed the test:
- Does not like SB16 emulation

Why the demo failed the test:
- Does not detect QEMU's display as VGA or PS/1 compatible

Why the demo failed the test:
- Missing graphics

Why the demo failed the test:
- Missing graphics

Why the demo failed the test:
- Missing graphics, scrolls way too fast

Why the demo failed the test:
- Does not like SB16 emulation

Why the demo failed the test:
- Disney Sound Source / LPT DAC output does not work

Why the demo failed the test:
- Disney Sound Source / LPT DAC output does not work

Why the demo failed the test:
- Disney Sound Source / LPT DAC output does not work

Why the demo failed the test:
- Does not support PC speaker "digitized speech"
Bochs will not be faulted for:
- Does not support LPT DAC

Why the demo failed the test:
- Does not render PC music
- Demo seems to hang when starting
Not QEMU's fault:
- Does not support LPT DAC

Why the demo failed the test:
- DOSBox SVN does not render the rapid changes to the PC speaker the demo uses (1ms precision limit)

Why the demo failed the test:
- DOSBox SVN does not render the rapid changes to the PC speaker the demo uses (1ms precision limit)

Why the demo failed the test:
- DOSBox SVN does not render the rapid changes to the PC speaker the demo uses (1ms precision limit)

Why the demo failed the test:
- VGA hpel is jumpy
- PC speaker music sounds erratic and deranged

Why the demo failed the test:
- Does not render PC speaker music

Do not select Sound Blaster, LPT DAC, or PC speaker at this time

Why the demo failed the test:
- Uses Direct DAC SB output, which QEMU apparently does not understand

Why the demo failed the test:
- Demo hangs on startup

Why the demo failed the test:
- Demo hangs after initial scroller
- Scroller exhibits odd "mode x" tearlines

Why the demo failed the test:
- Demo hangs after initial scroller

Why the demo failed the test:
- Detects SB16 emulation, and hangs

Why the demo failed the test:
- DOSBox Staging crashes with a segmentation fault

- Demo does not support SB16 emulation, but deals with it well enough

Why the demo failed the test:
- Detects SB16 emulation and hangs

Why the demo failed the test:
- Noticeable issues with closing credits and panning. Page flipping is visible and credits cut off at top.

Why the demo failed the test:
- Noticeable issues with closing credits and panning. Page flipping is visible and credits cut off at top.

Why the demo failed the test:
- Serious VGA panning issues
- Hangs after initial title

Why the demo failed the test:
- Hangs after initial title

Why the demo failed the test:
- Disney sound source / LPT DAC output doesn't work
- PC speaker output is distorted (compared to DOSBox-X)

Why the demo failed the test:
- Disney sound source / LPT DAC output doesn't work
- PC speaker output is distorted (compared to DOSBox-X)

Why the demo failed the test:
- Disney sound source / LPT DAC output doesn't work
- PC speaker output is distorted (compared to DOSBox-X)

Why the demo failed the test:
- Bochs doesn't support digitized PC speaker audio
Not Bochs fault:
- Does not support LPT DAC

Why the demo failed the test:
- QEMU does not support digitized PC speaker audio
Not QEMU's fault:
- Does not support LPT DAC

Why the demo failed the test:
- Does not support SB16 emulation
- VGA panning is broken

Why the demo failed the test:
- Hangs at Sound Blaster initialization

Why the demo failed the test:
- Runtime error

Why the demo failed the test:
- Hangs at Sound Blaster initialization

Why the demo failed the test:
- QEMU does not support Sound Blaster direct DAC output

Why the demo failed the test:
- QEMU does not support DSP commands used by this demo

Why the demo failed the test:
- Staging segfaults

Why the demo failed the test:
- Does not detect SB emulation

Why the demo failed the test:
- SB16 emulation not compatible with this demo

Why the demo failed the test:
- Not compatible with SB16 emulation

Why the demo failed the test:
- Direct DAC output not supported

Why the demo failed the test:
- Demo hangs on startup

Why the demo failed the test:
- QEMU does not support demo's use of Sound Blaster playback

Why the demo failed the test:
- "MOV_EwSw: can't use this segment register 6"

Why the demo failed the test:
- Does not work against SB16 emulation

Why the demo failed the test:
- Bochs does not emulate per-scanline effects

Why the demo failed the test:
- Hangs at startup message

Why the demo failed the test:
- Disney Sound Source / LPT DAC output doesn't work

Why the demo failed the test:
- Disney Sound Source / LPT DAC output doesn't work

Why the demo failed the test:
- Disney Sound Source / LPT DAC output doesn't work

Why the demo failed the test:
- Display flickering

Why the demo failed the test:
- Demo requires Sound Blaster, which Bochs is not emulating

Why the demo failed the test:
- Demo uses Sound Blaster DSP commands not recognized by QEMU

Why the demo failed the test:
- Bochs does not support per-scanline effects

Why the demo failed the test:
- QEMU does not support per-scanline effects

Why the demo failed the test:
- Nothing appears on screen

Why the demo failed the test:
- Nothing appears on screen

Why the demo failed the test:
- Not enough memory, COMMAND.COM error (?)

- Demo uses per-scanline palette changes mid-frame. There are two different palettes between the top logo and the bottom animation. Demo requires ET4000-specific hardware, therefore cannot use machine=vgaonly

- Demo uses per-scanline palette changes mid-frame. There are two different palettes between the top logo and the bottom animation. Demo requires ET4000-specific hardware, therefore cannot use machine=vgaonly

Why the demo failed the test:
- Runtime error

Why the demo failed the test:
- Runtime error

Why the demo failed the test:
- Runtime error

Why the demo failed the test:
- SB emulation not supported by QEMU. Even if you select "no sound"

Why the demo failed the test:
- It doesn't think QEMU provides a VGA display

Why the demo failed the test:
- QEMU Sound Blaster emulation is incompatible

Why the demo failed the test:
- Endless loop?

Why the demo failed the test:
- QEMU does not support Direct DAC SB output

Why the demo failed the test:
- VGA effects not emulated in Bochs

Why the demo failed the test:
- QEMU does not emulate SB direct DAC output

Why the demo failed the test:
- QEMU does not support SB Direct DAC output

Why the demo failed the test:
- VGA mode glitches during the introductory sequence

Why the demo failed the test:
- QEMU does not support Direct DAC SB output

Why the demo failed the test:
- QEMU causes Sound Blaster support to hang on startup

- Demo Sound Blaster support doesn't work in Bochs, but you can ask it not to use Sound Blaster

Why the demo failed the test:
- Sound Blaster output doesn't work in Bochs
- Flickering video problems in Bochs

Why the demo failed the test:
- QEMU does not support Direct DAC nor the sound playback when the fist punches through the monitor

It plays a sample, then that's it? Is that correct?

Why the demo failed the test:
- Hangs on startup, QEMU does not understand Sound Blaster detection

Demo will exit if you touch ANY key, including the key you last pressed during the prompts if it was still held down when it started.

Why the demo failed the test:

Why the demo failed the test:
- SVN does not support 'goldplay' mode

Why the demo failed the test:
- SVN does not emulate the goldplay mode

Why the demo failed the test:
- SVN version does not emulate goldplay method

Why the demo failed the test:
- SVN version does not emulate goldplay, DSP nagging, and the ability to sense the end of the DSP block through the DSP busy cycle

Why the demo failed the test:
- "Exit to error: Illegal GRP4 Call 4"

Why the demo failed the test:
- machine=svga_et4000 will not render per-scanline effects at startup correctly
- scroller is not rendered correctly, though background works correctly

Why the demo failed the test:
- machine=svga_et4000 will not render per-scanline effects at startup correctly
- scroller is not rendered correctly, though background works correctly

Why the demo failed the test:
- Does not support per-scanline VGA "warping" effects
- Second title animation is off-center
- Panning and scroller are completely messed up

Why the demo failed the test:
- Fails to show anything, exits to DOS
Notes:
- Sound Blaster output ACTUALLY WORKS WITH QEMU!

Why the demo failed the test:
- SVN version does not emulate DSP alias port 22Dh
- SVN version does not support backwards DMA

Why the demo failed the test:
- Planet "flickers" during main music playback in SVN

Why the demo failed the test:
- Exit to error: Illegal GRP4 Call 6

Why the demo failed the test:
- Crashes, spews random characters on screen

The demo does NOT work, but the following reasons do not count as a failure on SVN's part, because the demo would most likely fault the same on real hardware:
- DOSBox-SVN does not implement upper "reserved" bits in VGA status, which the demo requires in order not to hang

- Sound Blaster output will halt after about one second, because the demo does not unmask the IRQ and doesn't bother to. This is not DOSBox-SVN's fault, and does not count as a failure.

Why the demo failed the test:
- Crashes, invalid opcode errors

Why the demo failed the test:
- Demo prompts for sound card, then says that you need to run LEASTER.COM (I just did!)

Why the demo failed the test:
- Sound blaster support doesn't work (hangs)

Why the demo failed the test:
- SVN does not support "goldplay mode"

Why the demo failed the test:
- Disney Sound Source / LPT DAC output doesn't work

Why the demo failed the test:
- SVN version does not emulate goldplay method

Why the demo failed the test:
- SVN does not emulate "goldplay" mode
- Disney Sound Source / LPT DAC output doesn't work

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- Demo crashes on startup

Important notes:
- Sound Blaster output: DO NOT specify a sample rate higher than 20KHz. The demo will send Direct DAC output at the rate you specify. It will not wait for DSP ready, and it will not cap the sample rate within range. It will crash at higher than 25KHz sample rates.

Why the demo failed the test:
- Disney Sound Source / LPT DAC output doesn't work
Important notes:
- Sound Blaster output: DO NOT specify a sample rate higher than 20KHz. The demo will send Direct DAC output at the rate you specify. It will not wait for DSP ready, and it will not cap the sample rate within range. It will crash at higher than 25KHz sample rates.

Why the demo failed the test:
- Blanking overlap cuts off first few scanlines
- Panning down to skyline is not emulated correctly

Not a failure, but issues to investigate:
- Sound Blaster Pro sample rate difference between DOSBox-X and DOSBox-SVN

Why the demo failed the test:
- VGA palette flashing when some parts begin

Notes:
- VGA palette error between opening credits and animation (if cycles=20000)

Real hardware check:
- Some of the graphics flashing by... are those errors, artifacts, or accurate to real hardware?

Why the demo failed the test:
- Demo complains about EMS and exits, even though EMS has been turned off (lacks DOSBox-X fix where ems=false means not to set interrupt vector)

Real hardware check:
- I understand the intent is bright flashing colors, but, is the VGA palette animation correct?

Why the demo failed the test:
- LPT DAC output doesn't work. Selecting "Disney Sound Source" works.

Why the demo failed the test:
- SVN does not emulate "goldplay"
- LPT DAC output does not work

Why the demo failed the test:
- LPT DAC output does not work (Disney Sound Source output works)

Why the demo failed the test:
- LPT DAC output does not work
- SVN does not emulate "goldplay"

Why the demo failed the test:
- Code page conversion is preventing the demo from accessing it's mod file (file name "CRISIS" extension \xFF)

Real hardware testing:
- Toshiba 1997 laptop with SBPro compatible (OPL3-SAx): Crash on startup (two times, two tests)
- COMP-Old486y91, GUS, PAS: Exits normally, stops Sound Blaster output. Note the "Sound Blaster" in this case is a Pro Audio Spectrum.

Why the demo failed the test:
- LPT DAC output does not work
- Sound Blaster emulation is not precise enough to avoid occasional artifacts in Sound Blaster music playback
! PLEASE TEST ON REAL HARDWARE

Why the demo failed the test:
- LPT DAC output is *sometimes* very low fidelity
- "Atomic" logo flickers

Notes:
- LPT DAC output uses I/O port 3BC

Why the demo failed the test:
- SVN does not support "goldplay"

Why the demo failed the test:
- LPT DAC output does not work
- SVN does not emulate "goldplay"

Why the demo failed the test:
- LPT DAC output doesn't work
- SVN does not emulate "goldplay"

Why the demo failed the test:
- SVN does not emulate "goldplay"
- LPT DAC output doesn't work
The following is not DOSBox's fault and does not count as failure:
- Demo does not unmask IRQ

Why the demo failed the test:
- SVN does not support "goldplay"

Why the demo failed the test:
- LPT DAC output does not work (Disney Sound Source works)

Why the demo failed the test:
- Flickering during linear algebra part
- VGA palette errors, mid-demo text (should fade in from black but doesn't)

Why the demo failed the test:
- LPT DAC output doesn't work
The following is not DOSBox's fault and does not count as failure:
- Demo does not unmask IRQ, DOSBox SVN does not provide "unmask IRQ" for Sound Blaster emulation

Real hardware:
- Toshiba 1997 laptop: Crashes on startup
- COMP-Old486y91, GUS, PAS: Works with Pro Audio Spectrum, music is audible (The 486SX 33MHz is slow enough, apparently)

Why the demo failed the test:
- Demo does not respond to keyboard (masks IRQ1 and expects BIOS to unmask on INT 16h)

Why the demo failed the test:
- SVN does not support "goldplay"

Why the demo failed the test:
- SVN does not support "goldplay"

May not be DOSBox's fault:
- Hitting SPACEBAR at certain key points of the demo can cause a divide by zero exception

Hardware testing suggested:
- What VGA hardware does this work on (and where does it break)
Demo tricks:
- Per-scanline manipulation of color palette, VGA scanline width, display end

Why the demo failed the test:
- DOSBox SVN does not emulate undocumented INT 16h AH=1 behavior where the BIOS sets CF and ZF to specific values, which this demo requires to "sense" keyboard input. Such behavior has been confirmed with older (1992-ish) 386 BIOSes.

Notes:
- Demo does not stop GUS music and does not unhook music interrupt. Running anything after this point will crash DOS. Will mark as PASS if demo does the same thing on REAL HARDWARE.
Real hardware:
- Toshiba 1997 laptop (no GUS): Runs fine, exits fine
- COMP-Old486y91, GUS, PAS: Runs fine, does not respond do keyboard during credits screen.

Why the demo failed the test:
- DOSBox SVN does not emulate undocumented INT 16h AH=1 behavior where the BIOS sets CF and ZF to specific values, which this demo requires to "sense" keyboard input. Such behavior has been confirmed with older (1992-ish) 386 BIOSes.

Why the demo failed the test:
- SVN does not support "goldplay mode"
The following bugs are NOT DOSBox SVN's fault and do not count as failure:
- Demo does not unmask SB IRQ and doesn't bother to (Sound Blaster music stops after about one second).

Demo problems to examine:
- The 3D house part flickers badly if the cycle count is above a certain amount

Why the demo failed the test:
- SVN does not support "goldplay"

Why the demo failed the test:
- LPT DAC output does not work
- SVN does not emulate "goldplay"

Why the demo failed the test:
- SVN limits per-scanline VGA raster to machine=vgaonly. This demo requires machine=svga_et4000 to avoid graphics from appearing at the bottom when panning.

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- "Exit to error: PIC: 4 byte interval not handled"

Why the demo failed the test:
- Runtime error at startup
Real hardware testing:
- Toshiba Laptop 1997, MS-DOS 6.22: Works fine

Why the demo failed the test:
- Scroller part is broken (it doesn't scroll, just "judders")

Why the demo failed the test:
- Scroller part is broken (it doesn't scroll, just "judders")

Why the demo failed the test:
- SVN does not emulate GoldPlay

Why the demo failed the test:
- LPT DAC output doesn't work
- SVN does not support "goldplay"

Why the demo failed the test:
- LPT DAC output doesn't work
- SVN does not emulate "goldplay"

Why the demo failed the test:
- LPT DAC output doesn't work

Why the demo failed the test:
- "Exit to error: PIC: 4 byte interval not handled"

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- SVN does not emulate "goldplay"

- Demo calls BIOS to pause for a specific amount of time. BIOS call relies on RTC timer interrupt, which the demo has masked accidentally. DOSBox-SVN does not unmask the IRQ before waiting for the specific amount of time, therefore, demo hangs.

Why the demo failed the test:
- SVN does not emulate Goldplay

Why the demo failed the test:
- SVN does not emulate goldplay mode, IRQ stops

Notes:
- This demo seems to have very particular Sound Blaster DMA timing. There are very minor Sound Blaster artifacts during the demo.
! PLEASE RUN ON REAL HARDWARE

Why the demo failed the test:
- Sound Blaster audio st-stutters-rs for some reason

Why the demo failed the test:
- Demo crashes at one part, not enough memory

Why the demo failed the test:
- Demo runs but keyboard input on the sound config menu triggers too much of the menu and sound configuration is impossible

Why the demo failed the test:
- Demo hangs, don't run

Why the demo failed the test:
- VGA palette flashing between modesets
- SVN does not support "goldplay"

Notes:
- Gravis Ultrasound support does not work. Will mark PASSED if this occurs on real hardware.
Real hardware testing:
- COMP-Old486y91, GUS, PAS: Selected GUS, does not play music (GUS MAX card).

Notes:
- Demo is a bit unstable and may write random garbage into extended memory.
The following dosbox.conf settings are known to cause the demo to crash:
- ems=false
- memsize=8 to memsize=12

Notes:
- Demo palette handling seems to require Paradise/Western digital VGA DAC behavior, or else VGA palette is all wrong.
TODO:
- Run the demo on various SVGA chipsets to see where it works and where it fails. Perhaps the DOSBox-X implemented behavior for machine=svga_paradise is more common than thought.

Why the demo failed the test:
- VGA color palette problems

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- "Exit to error: PIC: 4 byte interval not handled"

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- VGA color palette flashing at certain points

Why the demo failed the test:
- Demo will randomly hang in certain parts where interrupt handler has the wrong data segment loaded, and ends up waiting for SB DSP ready from the DMA controller
- Sound Blaster popping and crackling

Why the demo failed the test:
- No sound

Why the demo failed the test:
- Uses protected mode. "Mapped vectors" eventually causes a divide by zero fault which is very poorly handled by the code.
Real hardware test:
- Toshiba Laptop, 1997, Pentium, MS-DOS 6.22: Crashed to DOS with an exception immediately
- COMP-Old486y91, GUS, PAS: Runs normally, does not crash (Something to do with 486SX/33MHz?). No sound. Exits to DOS normally.

Why the demo failed the test:
- Uses protected mode. "Mapped vectors" eventually causes a divide by zero fault which is very poorly handled by the code.
Real hardware test:
- Toshiba Laptop, 1997, Pentium, MS-DOS 6.22: Crashed to DOS with an exception immediately
- COMP-Old486y91, GUS, PAS: Runs normally, does not crash (Something to do with 486SX/33MHz?). No sound. Exits to DOS normally.

Why the demo failed the test:
- Demo doesn't GPF rapidly, but stops and does not respond to the keyboard
Real hardware test:
- Toshiba Laptop, 1997, Pentium, MS-DOS 6.22: Crashed to DOS with an exception immediately

Why the demo failed the test:
- Demo doesn't GPF rapidly, but stops and does not respond to the keyboard
Real hardware test:
- Toshiba Laptop, 1997, Pentium, MS-DOS 6.22: Crashed to DOS with an exception immediately

Why the demo failed the test:
- "interrupt(): soft_int && (gate.dpl < CPL)"

Why the demo failed the test:
- Hangs after initial "do you like scrollers"

Why the demo failed the test:
- SVN does not support "goldplay"

Why the demo failed the test:
- SVN does not support "goldplay"

Why the demo failed the test:
- Mode X VGA graphics come out blocky because SVN does not implement Sequencer alias registers

- One part terminates (INT 21h AH=4Ch) by zeroing it's code segment out from under itself, then
executing a MOV AH,4Ch + INT 21h from the prefetch queue.

Why the demo failed the test:
- Crashes, scribbles on ROM BIOS

Why the demo failed the test:
- Does not respond to keyboard (polls I/O port 64h with interrupts enabled, which means BIOS keyboard handler takes care of it before it can see it)

Why the demo failed the test:
- SVN does not emulate "goldplay"
The following does not count against DOSBox SVN:
- Demo does not unmask the IRQ, causing Sound Blaster playback to eventually stop

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- SVN does not emulate "goldplay"

- Demo crashes at the end. However in typhfix.zip, it's explained that
it did that at the party anyway, and that's why the fixed re-release,
therefore, it passes.

Notes:
- Sound Blaster Pro output crashes
Real hardware testing:
- COMP-Old486y91, GUS, PAS: - Sound Blaster Pro output crashes (even against PAS)

Why the demo failed the test:
- LPT DAC output doesn't work.
- SVN does not support Goldplay.

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- machine=svga_et4000 does not allow per-scanline effects as required by one part of the demo

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- Gravis Ultrasound output doesn't work

Notes:
- Does not exit cleanly to DOS, leaves stuck notes. However the music driver does not leave any interrupts hooked on exit. Perhaps this is how it would behave on real hardware.
Real hardware:
- COMP-Old486y91, GUS, PAS: Can't find ultrasound card for some reason (?)

Why the demo failed the test:
- GUS music sounds very sparse, usually one instrument over and over, which doesn't sound right (poke to addresses outside 1MB)
- "GET GOING" overdraws The Tick when it should scroll over with transparency

Notes:
- Demo runs, but exits to DOS leaving it's keyboard interrupt handler in place. System will crash when you hit a key (trigger IRQ1)
Real hardware:
- COMP-Old486y91, GUS, PAS: Demo runs, exits, leaves keyboard unresponsive.

Warning:
- This demo has rapid flickering (to "cross-blend" frames). Unless your monitor refresh rate happens to match the demo's the effect will look like glitching, but it isn't.

Why the demo failed the test:
- Demo is shown at half-height for some reason

Why the demo failed the test:
- SVN does not emulate "goldplay"
The following is not marked as failure against SVN:
- Does not offer option to unmask the SB IRQ

Why the demo failed the test:
- VGA palette errors

Why the demo failed the test:
- SVN does not emulate "goldplay"

Why the demo failed the test:
- SVN does not support "goldplay"

Why the demo failed the test:
- Lemmings rotozoomer is broken

Why the demo failed the test:
- Demo works, but, uses active display area effects in one part
(the part insulting Nintendo) which DOSBox-X does not render properly.

Notes:
- Crashes after multiple-plane checkboard effect
Real hardware:
- Toshiba Laptop 1997 Pentium: Same crash
- COMP-Old486y91, GUS, PAS: It made it past the checkerboard, to the circular cloud rendering (something about 486SX processors?), then crashed

Why the demo failed the test:
- Last demo part shows nothing on the screen. Debugging functions show that the palette is completely black and it tried to open an endpart.pal.
- When credits finish, screen remains blank, but you are at the DOS prompt.
Real hardware:
- Toshiba laptop 1997: Crashed, hung
- COMP-Old486y91, GUS, PAS: Demo runs normally, then crashes/hangs (with audible beeps from the 486SX unit).

Why the demo failed the test:
- Last demo part shows nothing on the screen. Debugging functions show that the palette is completely black and it tried to open an endpart.pal.
- When credits finish, screen remains blank, but you are at the DOS prompt.
Real hardware:
- Toshiba laptop 1997: Crashed, hung
- COMP-Old486y91, GUS, PAS: Demo runs normally, then crashes/hangs (with audible beeps from the 486SX unit).

Why the demo failed the test:
- Demo was unable to use the GUS emulation to play music

Why the demo failed the test:
- Demo didn't make it past first part

Why the demo failed the test:
- Last demo part shows nothing on the screen. Debugging functions show that the palette is completely black and it tried to open an endpart.pal.
- When credits finish, screen remains blank, but you are at the DOS prompt.
Real hardware:
- Toshiba laptop 1997: Crashed, hung
- COMP-Old486y91, GUS, PAS: Demo runs normally, then crashes/hangs (with audible beeps from the 486SX unit).

Why the demo failed the test:
- Last demo part shows nothing on the screen. Debugging functions show that the palette is completely black and it tried to open an endpart.pal.
- When credits finish, screen remains blank, but you are at the DOS prompt.
Real hardware:
- Toshiba laptop 1997: Crashed, hung
- COMP-Old486y91, GUS, PAS: Demo runs normally, then crashes/hangs (with audible beeps from the 486SX unit).

Why the demo failed the test:
- Demo was unable to use the GUS emulation to play music

Why the demo failed the test:
- Demo didn't make it past first part

Why the demo failed the test:
- SVN does not support "goldplay"

Why the demo failed the test:
- GUS playback never starts, demo never starts

Notes:
- Do not use Sound Blaster Pro, it crashes loading the credits part. This is acknowledged by the democoder in the readme.

Why the demo failed the test:
- Disney Sound Source / LPT DAC output doesn't work

Why the demo failed the test:
- Runtime error

Notes:
- GUS support does not work (crashes)
Real hardware:
- Toshiba laptop 1997, Yamaha OPL3-SAx: Crashes/hangs after initial logo
- COMP-Old486y91, GUS, PAS: Crashes/hangs after prompts for GUS playback (stuck in the same way as DOSBox-X)

Why the demo failed the test:
- SVN does not implement DSP alias port

Why the demo failed the test:
- Cannot keep SB IRQ going

Why do Beanix demos suck at playing sound? They seem to have consistent issues with
masking IRQs and then expecting the DOS or BIOS to unmask what is needed.

Why the demo failed the test:
- SVN does not support goldplay method

Demo bugs:
- When the demo probes the Sound Blaster IRQ, it uses a software timeout
(I/O delay) that can fail to detect the IRQ if it takes longer than
500ns for it's IRQ handler to receive the signal.
- If the demo cannot detect the Sound Blaster IRQ, then it will take
IRQ 0 (the timer interrupt) as the Sound Blaster IRQ, which will then
result in incorrect music playback.

Why the demo failed the test:
- Broken mode X graphics

Why the demo failed the test:
- SVN does not support Goldplay mode

Why the demo failed the test:
- It thinks there's an EMS manager present, because SVN does not offer zeroing INT 67h

Why the demo failed the test:
- Uses VGA active display animation (to open up at one demo part) which DOSBox-X does not properly support yet

Why the demo failed the test:
- Demo crashes midway through, with blank screen. Music keeps playing though. It does respond to hitting ESC on the keyboard to exit to DOS.
Real hardware:
- Toshiba Laptop 1997: Runs fine
- COMP-Old486y91, GUS, PAS: Demo hangs midway through, at rectangle and bouncing "ball", music keeps playing. It does however respond to the keyboard so it is possible to hit ESC to drop to DOS.

Notes:
- Demo refuses to talk to the Gravis Ultrasound.
Real hardware:
- Toshiba Laptop 1997 Chips & Tech 65554, Yamaha OPL3-SAx (Sound Blaster Pro): Runs fine, but then kills the VGA output
- COMP-Old486y91, GUS, PAS: Refuses Gravis Ultrsound, just as it does in DOSBox-X

Why the demo failed the test:
- Can't quite do the plasma effect

Why the demo failed the test:
- SVN does not support Goldplay method

Why the demo failed the test:
- Mode X output garbled

- Random I/O ports are being accessed during demo. The GUS music support routines have a known bug in the ISR that pops DS and then attempts to read the GUS base with whatever DS segment happened to exist. The way the ISR is written implies the same issue on real hardware.

Why the demo failed the test:
- SVN does not support Goldplay

Why the demo failed the test:
- Demo does not respond to keyboard

Why the demo failed the test:
- GUS music "hung" mid-demo

Why the demo failed the test:
- SVN does not support Goldplay

Demo notes:
- Requires GUS DMA on IRQ 1, or it does not work.
- Music sounds awful, out of tune, off key, like there's something wrong with the tracker or the music data. It has been confirmed that the music sounds just as bad on real Gravis Ultrasound MAX hardware.
- For some reason, it sounds even worse when run in DOSBox SVN.

Demo notes:
- Requires GUS DMA on IRQ 1, or it does not work.
- Music sounds awful, out of tune, off key, like there's something wrong with the tracker or the music data. It has been confirmed that the music sounds just as bad on real Gravis Ultrasound MAX hardware.
- For some reason, it sounds even worse when run in DOSBox SVN.

- Does not run at all

- Does not run at all

Why the demo failed the test:
- Always crashes at one point with divide by zero.
Real hardware:
- Toshiba Laptop 1997, Chips & Tech 65554 SVGA. Palette errors during "This demo is compatible with Tseng/Trident", render errors during animation. The ship gets garbled at one part in the same way that divide by zero crashes in DOSBox-X.
- COMP-Old486y91, GUS, PAS, ET4000: Some flicker during 3D maze (as if bad page flipping). GUS music works (IRQ 11, DMA 5). Does not crash during space part (something about a 486SX/33MHz?)

Notes:
- GUS playback routine has the common bug that causes random I/O ports to be accessed during the demo. A patch is applied to the EXE to fix this.
- The demo installs a keyboard ISR, but the ISR gets corrupted at some point, and the demo does not restore the original ISR on exit. A TSR is run to fix this.

Notes:
- GUS playback routine has the common bug that causes random I/O ports to be accessed during the demo. A patch is applied to the EXE to fix this.
- The demo installs a keyboard ISR, but the ISR gets corrupted at some point, and the demo does not restore the original ISR on exit. A TSR is run to fix this.

Why the demo failed the test:
- Gravis Ultrasound music plays way too fast. You may experience this demo better using the Sound Blaster mode.

Why the demo failed the test:
- Demo exits to DOS, but then hitting a key causes a crash.
Notes:
- GUS music playback (or the demo itself?) seems to read and write random I/O ports during the demo. DOSLIB utility to patch out that bug is available.

Why the demo failed the test:
- Demo exits to DOS, but then hitting a key causes a crash.
Notes:
- GUS music playback (or the demo itself?) seems to read and write random I/O ports during the demo. DOSLIB utility to patch out that bug is available.